The data were subjected to a cluster analysis to determine if respondents could be grouped into clusters based on their responses to the ten brand attitude Likert items for Tylenol.  The clusters were then analyzed to determine if the clusters were differentiated on the basis of sex.


Tylenol…

Two-means cluster analysis

Three-means cluster analysis

Cluster 1

Cluster 2

Cluster 1

Cluster 2

Cluster 3

Member count

26

100

6

36

84

…is a good brand

3.3 (0.9)

4.3 (0.4)

3.5 (0.8)

3.6 (0.9)

4.3 (0.4)

…works fast

2.7 (0.8)

3.9 (0.6)

2.3 (1.0)

3.3 (0.8)

4.0 (0.6)

…prevents heart attacks

2.4 (0.8)

2.8 (0.8)

2.2 (1.0)

2.6 (0.7)

2.8 (0.8)

…is safer

2.8 (0.7)

3.2 (0.7)

3.2 (0.8)

2.8 (0.7)

3.2 (0.7)

…is trustworthy

2.7 (0.7)

4.1 (0.6)

3.3 (0.8)

3.0 (0.8)

4.2 (0.6)

…would be prescribed by my doctor

3.5 (0.8)

3.7 (0.7)

4.0 (0.6)

3.4 (0.7)

3.7 (0.7)

…is easy to use

3.5 (0.9)

4.1 (0.7)

4.3 (0.5)

3.2 (0.9)

4.3 (0.5)

…doesn’t have too many side effects

3.0 (0.8)

3.8 (0.7)

3.8 (0.4)

2.9 (0.6)

4.0 (0.6)

…is made by a company that cares about my health

2.8 (0.9)

2.3 (0.7)

2.3 (1.2)

3.0 (0.7)

3.3 (0.8)

…provides effective pain relief

3.0 (0.9)

4.0 (0.4)

2.0 (0.6)

3.6 (0.7)

4.1 (0.4)

Mean (Standard deviation)

In the two-means cluster analysis, the mean Likert score is higher in cluster 2 than cluster 1 for all attributes except “Tylenol is made by a company that cares about my health.”  The pattern for the three-means cluster analysis is much less neat.

Table 1: Analysis of variance for three-means cluster analysis

 

Cluster

Error

F

Mean square

Degrees of freedom

Mean square

Degrees of freedom

…is a good brand

6.5

2

0.4

123

17.04*

…works fast

12.2

2

0.5

123

25.37*

…prevents heart attacks

1.0

2

0.6

123

1.68

…is safer

2.0

2

0.5

123

4.04*

…is trustworthy

21.0

2

0.4

123

50.06*

…would be prescribed by my doctor

1.7

2

0.5

123

3.16*

…is easy to use

15.6

2

0.4

123

36.89*

…doesn’t have too many side effects

15.0

2

0.4

123

38.15*

…is made by a company that cares about my health

3.3

2

0.6

123

5.73*

…provides effective pain relief

15.1

2

0.3

123

54.02*

* p ≤ 0.15

Table 2: Analysis of variance for two-means cluster analysis

 

Cluster

Error

F

Mean square

Degrees of freedom

Mean square

Degrees of freedom

…is a good brand

19.85

1

0.32

124

61.74*

…works fast

29.68

1

0.43

124

68.62*

…prevents heart attacks

3.07

1

0.60

124

5.15*

…is safer

4.01

1

0.51

124

7.86*

…is trustworthy

42.65

1

0.41

124

104.01*

…would be prescribed by my doctor

0.47

1

0.55

124

0.87

…is easy to use

8.45

1

0.60

124

14.06*

…doesn’t have too many side effects

12.60

1

0.53

124

23.78*

…is made by a company that cares about my health

4.22

1

0.59

124

7.14*

…provides effective pain relief

24.52

1

0.32

124

75.64*

* p ≤ 0.15

 

Male

Female

Cluster 1
Row %
Column %
Total %

14
60.9%
28.6%
11.5%

9
39.1%
12.3%
7.4%

Cluster 2
Row %
Column %
Total %

35
35.4%
71.4%
28.7%

64
64.6%
87.7%
52.5%

 

c2 = 5.06*

* p ≤ 0.15

In 85 samples out of 100 samples drawn from the same population as this sample, the cross-tabulation displayed above would be about what it is here.  The results of this test can be projected to the population.